Building High Availability Windows Server 2003 Solutions

by Shapiro, Jeffrey R.; Policht, Marcin

• ISBN13: 9780321228789
• ISBN10: 0321228782
• Edition: 1st
• Format: Paperback
• Copyright: 2004-12-14
• Publisher: Addison-Wesley Professional

Summary

"This is the definitive resource for every network administrator, consultant, and architect who needs to maximize availability, scalability, and performance in Windows server environments." "Shapiro and Policht offer a clear, concise roadmap for keeping Windows servers running 24x7 and delivering on even the most challenging service-level agreement. They provide real-world case studies and easy-to-use instruction designed to help readers make better decisions more rapidly."--BOOK JACKET.

Author Biography

Building High Availability Windows Server™ 2003 Solutions About the Authors

Jeffrey R. Shapiro has worked in Information Technology for nearly 15 years. He is an industry-celebrated author and has published more than a dozen books on IT, network administration, and software development. Jeffrey has written for numerous publications over the years as well. He also regularly speaks at events and frequently participates in training courses on Microsoft systems.

Jeffrey has specialized in Microsoft technologies since 1989. From 1992 to 1998, he was CTO for a leading software development company specializing in telephony solutions for business and was credited with designing the architecture for one of the first Windows-based computer telephony platforms.

In early 2003 Jeffrey was selected to lead the Novell NetWare to Windows Server 2003 migration project for Broward County, Florida. His mandate was to design the architecture for an Active Directory network that would replace the hundreds of servers and Novell Directory Services (NDS) required to support more than 80 agencies. He was also in charge of designing the architecture for three mission-critical, high availability, high-performance data centers supporting thousands of public servants in one of the largest population centers in the United States.

In late 2004 Jeffrey turned his attention almost exclusively to systems and software architecture. He recently formed Normal Data, Inc., a company that specializes in architecting software for enterprise information technology solutions http://www.codetimes.com. Jeffrey can be reached on the Web at [e-mail].

Marcin Policht has diverse experience in areas of scripting and programming, as well as system engineering and administration of large-scale, high availability, Windows-based environments. He has shared his expertise as a technical trainer and as a writer, authoring a number of books and Web articles on subjects varying from WMI scripting to Active Directory management.

© Copyright Pearson Education. All rights reserved.

Table of Contents

Acknowledgments	xvi
About the Authors	xviii
Preface	xix
PART I HIGH-PERFORMANCE WINDOWS COMPUTING
Chapter 1 The World of High-Performance, High Availability Windows Computing	3	(22)
Introduction	3	(1)
Service Level	4	(2)
Availability	6	(4)
High Availability, Downtime, and Failure	10	(8)
Scale-Out Availability and Windows Server 2003	14	(1)
Clustering	15	(1)
Scale-Up Availability	15	(1)
Scale-Out or Scale-Up'?	16	(1)
Share Everything Versus Share Nothing	17	(1)
High-Performance Computing	18	(5)
The Need for High-Performance Computing	18	(1)
High-Performance Computing for Everyone	19	(1)
Supercomputers in Every Closet	20	(1)
Processing and Memory	21	(1)
High-Performance Components	21	(1)
Microsoft and the Cornell Theory Center	22	(1)
Time-Out	23	(2)
Chapter 2 Choosing High-Performance Hardware	25	(10)
Introduction	25	(1)
Standards, Vendors, and Common Sense	26	(2)
Vendors	26	(1)
Common Sense	27	(1)
Choosing the CPU	28	(2)
Memory	30	(4)
DRAM	31	(1)
DRAM with EDO	32	(1)
Synchronous DRAM	32	(1)
Direct Rambus DRAM (RDRAM)	33	(1)
Time-Out	34	(1)
Chapter 3 Storage for Highly Available Systems	35	(36)
Introduction	35	(1)
Redundancy and Availability of Storage	36	(5)
RAID Refresher	41	(10)
RAID 1	44	(2)
RAID 5	46	(1)
RAID 10	46	(1)
RAID Controllers	47	(4)
Server Attached Storage Solutions	51	(2)
Network Attached Storage Solutions (NAS)	53	(4)
Storage Area Networks (SAN)	57	(7)
IP-Based Storage Solutions	64	(5)
Time-Out	69	(2)
Chapter 4 Highly Available Networks	71	(26)
Introduction	71	(1)
Backbone Design for High Availability	71	(3)
Bandwidth Field Notes	74	(11)
Ethernet	74	(3)
What to Look for in Network Interface Cards	77	(2)
Hubs, Switches, and Routers	79	(1)
Layer 2 Switches	80	(3)
Layer 3, Layer 4, and Beyond	83	(1)
Routers and Routing in High Availability Architecture	83	(1)
Using Hubs for Failover Interconnects	84	(1)
SAN Topology Primer	85	(8)
Fibre Channel	86	(2)
SAN Topology	88	(1)
Ports	88	(1)
Point-to-Point Topology	89	(1)
FC-AL	89	(2)
Fabric	91	(1)
Zoning	92	(1)
Architecting SAN Topology for High Availability	93	(2)
Time-Out	95	(2)
Chapter 5 Preparing the Platform for a High-Performance Network	97	(66)
Introduction	97	(1)
Architecting Primer	98	(7)
Create a Design Plan	100	(1)
Design Goals	100	(1)
Design Components	101	(1)
Design Decisions	102	(2)
Design Implications	104	(1)
Active Directory Services, Logical Architecture	105	(2)
Forest Plan for Highly Available Systems	107	(3)
Single Global Catalog	109	(1)
Domain Namespace	110	(28)
External DNS Domain Name	112	(1)
Domain Controllers (DCs)	112	(1)
Multi-Master Operations (Global Catalogs)	113	(1)
Single Master Operations (FSMO Roles)	114	(1)
Schema Master	115	(1)
Domain Naming Master	115	(1)
RID (Relative Identifier Master)	115	(1)
Primary Domain Controller Emulator	116	(1)
Infrastructure Master	116	(3)
Miscellaneous Roles for Domain Controllers	119	(1)
Preferred Group Policy Administrator Domain Controller	119	(1)
Time Service	119	(1)
Organizational Units	119	(5)
Group Policy Backgrounder	124	(4)
Password Policy	128	(8)
Event Log	136	(1)
Group Policy Objects for Cluster Servers	137	(1)
Active Directory Physical Architecture	138	(12)
Subnets	138	(5)
Site Links	143	(1)
Cost	144	(2)
Replication Schedule and Notification	146	(1)
Transports	147	(1)
Connection Objects	148	(1)
Site Link Bridge	149	(1)
Site Layout and Topology	149	(1)
AD Integrated DDNS (Dynamic DNS)	150	(4)
DNS Architecture	151	(1)
Hub Sites	152	(1)
Administration of DNS Servers	153	(1)
DDNS Configuration	153	(1)
WINS	154	(2)
Hub Sites	155	(1)
Administration of WINS Servers	156	(1)
DHCP (Dynamic Host Configuration Protocol)	156	(2)
DHCP Architecture	157	(1)
DHCP Parameters	158	(1)
Scope Details	158	(1)
Naming Conventions	158	(3)
Time-Out	161	(2)
Chapter 6 Building the Foundations for a Highly Available Architecture	163	(54)
Introduction	163	(1)
Windows Clustering 101	164	(9)
The Cluster Model	164	(7)
The Quorum Resource	171	(1)
Deployment Scenarios	172	(1)
Forest Creation Process	173	(3)
Installation of Support Server	174	(2)
Installation	176	(1)
Installation of Root Domain	177	(6)
Process	177	(6)
Quality Assurance	183	(1)
Forest Preparation, DNS, and Exchange	184	(2)
Installation of Bridgehead Servers and the Child Domain	186	(8)
Installing DHCP and WINS Services	194	(2)
Patching and Updating Domain Controllers	196	(1)
Exchange Domain Preparation	197	(1)
Creation of Initial Service and Administration Resources	198	(2)
Clustering	200	(13)
Create Shared Disk Resources	200	(2)
Prepare the Cluster Network	202	(1)
Start Server Cluster Wizard	202	(9)
Troubleshooting	211	(2)
Time-Out	213	(4)
PART II BUILDING HIGH AVAILABILITY WINDOWS SERVER 2003 SOLUTIONS
Chapter 7 High-Performance Print-Server Solutions	217	(8)
Introduction	217	(1)
Design Specifications	218	(4)
Installation	222	(2)
Install Spooler Resources	223	(1)
Time-Out	224	(1)
Chapter 8 High-Performance File-Server Solution	225	(20)
Introduction	225	(1)
Scale-Out Versus Scale-Up with File Servers	225	(3)
Design	228	(8)
Develop Lab Systems	230	(1)
Configure Hardware	231	(1)
Configure 2-Node Cluster Services	231	(1)
Deploy Standard File System Configuration	231	(1)
Define and Implement Backup/Restore Procedures	231	(1)
Create a File Server Security Plan	232	(1)
Configure Root of a Domain DFS	232	(1)
Set Up File Server Administration Tools	232	(1)
Define and Implement File Server Antivirus Strategy	232	(1)
General Configuration	232	(2)
Configuration for File Server Clusters	234	(2)
Installation	236	(2)
Standard File Share	236	(1)
Share or Hide Subdirectories	237	(1)
Installing the File Share Resource	237	(1)
High Availability Using Replication and Domain DFS	238	(5)
Time-Out	243	(2)
Chapter 9 High Availability, High-Performance SQL Server Solutions	245	(52)
Introduction	245	(1)
Scale-Out Versus Scale-Up with Microsoft SQL Server	246	(1)
Design	247	(3)
Failover for SQL Server	250	(1)
SQL Server Cluster Design Specs	251	(12)
Documenting the Dependencies	251	(1)
Understanding SQL Server Active/Passive Configurations	252	(1)
Active/Active Configurations and Multiple Instances	252	(2)
N+1 Configurations	254	(2)
Physical Disks	256	(3)
Memory	259	(1)
Local Disks	260	(2)
Standby Services-Advantages and Disadvantages	262	(1)
Clustering SQL Server	263	(7)
High Availability, High-Performance Notes	270	(5)
Storage Notes	270	(2)
Failover Resources	272	(1)
Enterprise Manager	273	(1)
Transactions and Logs	274	(1)
Configuration and Planning	275	(2)
The Role of Replication	277	(2)
Disaster Recovery	279	(1)
HA for Analysis Services (OLAP)	280	(12)
Clustering Analysis Services	281	(2)
Create Domain OLAP Administrators Group	283	(8)
Clustering SQL Server 2000 Analysis Services Troubleshooting and Best Practices	291	(1)
Troubleshooting, Maintenance, and Best Practices	292	(3)
Fragmentation	293	(1)
Operating System Level-Backup Utilities	293	(1)
Anti-Virus Software	294	(1)
Windows Updates	294	(1)
MBSA	294	(1)
Time-Out	295	(2)
Chapter 10 High Availability, High-Performance Exchange	297	(38)
Introduction	297	(2)
Scale-Out Versus Scale-Up with Microsoft Exchange	299	(1)
Design	300	(8)
Storage Group Architecture	304	(3)
Transaction Log Files	307	(1)
SMTP Queue Directory	307	(1)
Exchange Permissions in the Clustering Architecture	308	(1)
Getting Started with Exchange 2003 Clustering	309	(25)
Installing Exchange on the Cluster Nodes	309	(5)
The Exchange Virtual Server	314	(2)
Cluster Groups	316	(1)
Cluster Configurations	317	(3)
IP Addresses and Network Names	320	(1)
Creating the MSDTC Group	321	(1)
Creating the EVS	321	(7)
Creating an Exchange 2003 System Attendant Resource	328	(5)
Configuring a Clustered Back-End Server	333	(1)
Time-Out	334	(1)
Chapter 11 Load Balancing	335	(38)
Introduction	335	(1)
Scale-Out Revisited	336	(1)
Fault Tolerance and High Availability of NLB	337	(1)
Load Balancing for High Performance	338	(5)
Sharing Server Load	338	(1)
Virtual Servers	339	(1)
What Cannot Be Scaled	340	(2)
Selecting NLB Clustering Candidates	342	(1)
Network Load Balancing Architecture	343	(3)
Designing the NLB Cluster	346	(21)
Design Specifications	346	(7)
Port Rules	353	(3)
Setup and Configuration of the NLB Cluster	356	(5)
Example NLB Cluster: IIS	361	(1)
Example NLB Cluster: Terminal Services	361	(3)
Load Balancing and COM Application Servers	364	(2)
Multi-Tiered Server Farms	366	(1)
NLB Cluster Management	367	(3)
Administering the NLB Cluster	368	(1)
Troubleshooting	369	(1)
Disaster Recovery	370	(1)
Time-Out	370	(3)
Chapter 12 Internet Information Server	373	(28)
Introduction	373	(1)
IIS 6.0 and the Dedicated Web Server	373	(7)
Scale-Out Versus Scale-Up IIS	380	(5)
Round Robin DNS	382	(1)
Load Balancing	382	(3)
NLB for IIS	385	(1)
Planning and Configuration	386	(11)
IIS Storage	392	(2)
FTP Service	394	(3)
Troubleshooting	397	(1)
Maintaining the IIS Server Cluster	398	(1)
Disaster Recovery	399	(1)
Best Practices	399	(1)
Time-Out	400	(1)
Chapter 13 Looking for Trouble: Setting Up Performance Monitoring and Alerts	401	(58)
Introduction	401	(1)
Understanding the Windows Server 2003 Monitoring Systems	402	(22)
Event Viewer	404	(3)
Exploring System and Performance Monitoring Objects	407	(1)
Rate and Throughput	408	(1)
Understanding the Work Queue	408	(1)
Response Time	409	(1)
How Performance Objects Work	409	(1)
System Monitoring Tools	410	(1)
Working with the Performance Console and the System Monitor	411	(1)
How to Use System Monitor	412	(3)
Performance Logs and Alerts	415	(1)
Using Logs and Alerts	416	(2)
Monitoring the Servers	418	(1)
Monitoring for Bottlenecks	419	(2)
Understanding Your Server's Workload	421	(2)
Performance Monitoring Tips	423	(1)
Microsoft Operations Manager	424	(3)
MOM Rapid Fire Deployment	427	(29)
Verifying Software and Hardware Requirements	429	(1)
MOM Service Accounts	430	(1)
MOM Database Sizing	430	(2)
Design	432	(1)
SQL Server Notes	433	(1)
Installing MOM Databases	434	(6)
Installing the First Management Server	440	(2)
Installing the MOM Administrator and MOM Operator Consoles	442	(1)
Discovering Computers and Deploying Agents	442	(2)
Agent Failover	444	(1)
Installing System Center 2005 Reporting	445	(1)
Importing MOM 2005 Management Packs	446	(2)
Management Pack Management	448	(8)
Time-Out	456	(3)
Index	459

Excerpts

Building High Availability Windows Server 2003 Solutions Building High Availability Windows Server 2003 Solutions Preface The year 2004 will long be remembered as the year that saw the beginning of a huge push by companies and government organizations to once and for all migrate to a Windows Server operating system underpinned by Active Directory and Windows Server 2003. What is also significant about this year is that it will be remembered as the year Microsoft finally ended all support for Windows NT 4.0, the grandfather of the current version of Windows Server that many IT professionals now regard as the Serengeti of the operating system jungle. In 2004, many companies have finally made the move to ditch Novell NetWare. However, it is not simply enough to trade one operating system for another. Many IT shops going to Windows Server 2003 need to install and configure high availability, high-performance Windows Server 2003 systems that can service their needs day in, day out, 365 days a year. At the same time, they are also striving to lower the cost of installing, operating, and maintaining these systems and the overall cost of ownership (TCO). Windows Server 2003 delivers on all these points. As companies migrate to the platform that is thede factowinner in the network and operating system wars, they face a huge learning curve and dilemma on how best to set up high-performance Windows Server networks for maximum availability and power. Their aspirations come down to one thing: service level--"How do we do it with Windows Server 2003?" Companies that have made the decision to migrate to Windows Server 2003 ask how they can keep systems up 24/7 or how they can achieve three, four, and perhaps even five times the availability with Microsoft technology. Network administrators ask, "Do we cluster, do we load balance, do we do both, do we invest in hot standbys, replication...what works?" This book gives you the answers to those questions. It will also go further than just failover and fault tolerance and discuss monitoring and operations management and choosing the right technology to accompany Microsoft's high-performance and high availability offerings. This is the book that caters to your needs. It is about achieving service level and keeping systems up 24/7 with the Windows Server 2003 platform. This book provides a clear and concise roadmap for how to go about using Microsoft Server 2003, (in some cases) with third-party add-ons, for scalability, uptime, performance, and management--and for how to avoid trouble at the same time. Many administrators and engineers find it hard to make decisions about what they need to do. They hear that clustering and using load balancing is a black art--extremely difficult and prone to disaster. Up until today, their only resources for architecting a high availability solution has been rare and expensive consultants and overzealous consulting services engineers, particularly from hardware vendors. If you are turned to Microsoft technology to achieve your SLA, this book will be the foundation to turn to, to bring it all together. Microsoft now offers a rich toolset for administration and monitoring, not only what is built into the server products, but also with collateral offerings such as Microsoft Operations Manager (MOM) and Systems Management Server. According to Gartner, Microsoft will own the systems administration market, and possibly surge ahead of IBM, in the coming years. Efforts in this area became very evident in 2003 and 2004 with the advent of new versions of MOM and Systems Management Server. We have thus devoted an entire chapter to monitoring and installing MOM as the essential operations platform for any high availability network. The book is divided into two parts: Part I,